Application of artificial neural network and dynamic fault tree analysis to enhance reliability in predictive ship machinery health condition monitoring

Daya, Abdullahi and Lazakis, Iraklis (2021) Application of artificial neural network and dynamic fault tree analysis to enhance reliability in predictive ship machinery health condition monitoring. In: 2nd International Conference on Ship and Marine Technology, 2021-09-16 - 2021-09-17.

[thumbnail of Daya-Lazakis-GMO-SHIPMAR-2021-Application-of-artificial-neural-network-and-dynamic-fault-tree]
Preview
Text. Filename: Daya_Lazakis_GMO_SHIPMAR_2021_Application_of_artificial_neural_network_and_dynamic_fault_tree.pdf
Final Published Version

Download (1MB)| Preview

    Abstract

    The electric power generation system of most ships is powered by a group of diesel generators generally with redundancy to accommodate peak load periods or critical situations. Blackouts onboard ships portents a potential danger to navigation as well as the security and safety of the ship. Thus, understanding the factors affecting the reliability of individual diesel generators and the most critical component to failure is key to ensuring reliable performance of the generators. Therefore, this study was conducted on diesel power generation plant consisting of four Marine Diesel Generators onboard an Offshore Patrol Vessel (OPV). Findings indicates relatively low reliability, of less than 60 per cent within the first 24 months of the 78 operational months data analysed. Similarly, reliability importance measures were adopted to identify Critical components which contribute at least 40 per cent of failures on the sub systems of the diesel generators. The use of dynamic spare gates in the dynamic fault tree analysis has highlighted possible improvements through maintenance action or use of sensors to improve sub-system as well as individual diesel generator’s reliability. Additionally, Artificial Neural Networks classification using unsupervised learning was conducted to identify patterns in the data that signifies the onset of performance degradation in the diesel generators.

    ORCID iDs

    Daya, Abdullahi and Lazakis, Iraklis ORCID logoORCID: https://orcid.org/0000-0002-6130-9410;